{"id":"W2113944588","doi":"10.1093/bioinformatics/btn433","title":"Accurate sequence-based prediction of catalytic residues","year":2008,"lang":"en","type":"article","venue":"Bioinformatics","topic":"Machine Learning in Bioinformatics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":85,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta","funders":"Nankai University","keywords":"Support vector machine; Curse of dimensionality; Sequence (biology); Computer science; Classifier (UML); Protein structure prediction; Feature selection; Artificial intelligence; Biological system; Data mining; Pattern recognition (psychology); Protein structure; Chemistry; Biology; Biochemistry","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0012644,0.0009282973,0.0007739869,0.001279155,0.0002663598,0.0008248955,0.0006876886,0.000984329,0.00127031],"category_scores_gemma":[0.003775498,0.0002646203,0.0004586986,0.0008338963,0.0002762785,0.001141909,0.0004813195,0.0009375165,0.00190982],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003408814,"about_ca_system_score_gemma":0.0005245066,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001001528,"about_ca_topic_score_gemma":0.0009817609,"domain_scores_codex":[0.99914,0.0001581204,0.00008155651,0.000251814,0.0003145788,0.00005385392],"domain_scores_gemma":[0.9973908,0.001025018,0.000513437,0.0002741356,0.000697553,0.00009906089],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0008899393,0.0004321153,0.09294408,0.001790129,0.0002540501,0.0008233608,0.00009134696,0.1870845,0.2226762,0.0020636,0.01197492,0.4789757],"study_design_scores_gemma":[0.00004775807,0.0002276023,0.02225304,0.00008687928,0.00008470315,0.0008967205,0.00004387323,0.8575967,0.1083416,0.003860252,0.006517851,0.00004292096],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.4861636,0.008682016,0.4890255,0.0005475204,0.0001817484,0.0001137328,0.004077362,0.006864098,0.004344356],"genre_scores_gemma":[0.8359466,0.001578549,0.1518073,0.0001607575,0.0001211217,0.00006311066,0.008609102,0.0001872746,0.001526286],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.001279155,"threshold_uncertainty_score":0.006686807,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02821141345171289,"score_gpt":0.2624030186154391,"score_spread":0.2341916051637262,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}